goacc/reduction-2.f95 failure

goacc/reduction-2.f95 failure

[Steve Kargl]

Cesar, Thomas,

I just got a failure in the gfortran testsuite.

FAIL: gfortran.dg/goacc/reduction-2.f95   -O   scan-tree-dump-times \
gimple "target oacc_kernels map.tofrom:a .len: 4.." 1

Is this a known failure or expected?

reduction-2.f95 was committed yesterday.  Is something missing?

-- 
Steve

Re: goacc/reduction-2.f95 failure

[Cesar Philippidis]

On 12/02/2015 12:33 PM, Steve Kargl wrote:

> I just got a failure in the gfortran testsuite.
> 
> FAIL: gfortran.dg/goacc/reduction-2.f95   -O   scan-tree-dump-times \
> gimple "target oacc_kernels map.tofrom:a .len: 4.." 1
> 
> Is this a known failure or expected?
> 
> reduction-2.f95 was committed yesterday.  Is something missing?

How old is your build?

I think Tom tweaked the default data mapping for kernels variables
yesterday either right before or after I committed a patch which
introduced that test case. That should be fixed in this patch

  https://gcc.gnu.org/ml/gcc-patches/2015-12/msg00349.html

which applied I applied to trunk about an hour ago.

Cesar

Re: goacc/reduction-2.f95 failure

[Steve Kargl]

On Wed, Dec 02, 2015 at 12:42:40PM -0800, Cesar Philippidis wrote:
> On 12/02/2015 12:33 PM, Steve Kargl wrote:
> 
> > I just got a failure in the gfortran testsuite.
> > 
> > FAIL: gfortran.dg/goacc/reduction-2.f95   -O   scan-tree-dump-times \
> > gimple "target oacc_kernels map.tofrom:a .len: 4.." 1
> > 
> > Is this a known failure or expected?
> > 
> > reduction-2.f95 was committed yesterday.  Is something missing?
> 
> How old is your build?
> 
> I think Tom tweaked the default data mapping for kernels variables
> yesterday either right before or after I committed a patch which
> introduced that test case. That should be fixed in this patch
> 
>   https://gcc.gnu.org/ml/gcc-patches/2015-12/msg00349.html
> 
> which applied I applied to trunk about an hour ago.
> 

It's about an hour old.  I was testing the coarray events patch.
In preparation for the commit, I updated the gcc/testsuite directory
to get an up-to-date ChangeLog.  I probably picked up a part of 
a commit.  Rebuild from a top-of-tree as of 10 minutes ago.  I know
shortly if it is a false alarm.

-- 
Steve

Re: goacc/reduction-2.f95 failure

[Steve Kargl]

On Wed, Dec 02, 2015 at 12:49:24PM -0800, Steve Kargl wrote:
> On Wed, Dec 02, 2015 at 12:42:40PM -0800, Cesar Philippidis wrote:
> > On 12/02/2015 12:33 PM, Steve Kargl wrote:
> > 
> > > I just got a failure in the gfortran testsuite.
> > > 
> > > FAIL: gfortran.dg/goacc/reduction-2.f95   -O   scan-tree-dump-times \
> > > gimple "target oacc_kernels map.tofrom:a .len: 4.." 1
> > > 
> > > Is this a known failure or expected?
> > > 
> > > reduction-2.f95 was committed yesterday.  Is something missing?
> > 
> > How old is your build?
> > 
> > I think Tom tweaked the default data mapping for kernels variables
> > yesterday either right before or after I committed a patch which
> > introduced that test case. That should be fixed in this patch
> > 
> >   https://gcc.gnu.org/ml/gcc-patches/2015-12/msg00349.html
> > 
> > which applied I applied to trunk about an hour ago.
> > 
> 
> It's about an hour old.  I was testing the coarray events patch.
> In preparation for the commit, I updated the gcc/testsuite directory
> to get an up-to-date ChangeLog.  I probably picked up a part of 
> a commit.  Rebuild from a top-of-tree as of 10 minutes ago.  I know
> shortly if it is a false alarm.

Following up: it appears I had an inconsistent tree, so yes, a false alarm.

-- 
Steve